[PATCH] swsusp: write speedup

Switch the swsusp writeout code from 4k-at-a-time to 4MB-at-a-time.

Crufty old PIII testbox:
	12.9 MB/s -> 20.9 MB/s

Sony Vaio:
	14.7 MB/s -> 26.5 MB/s

The implementation is crude.  A better one would use larger BIOs, but wouldn't
gain any performance.

The memcpys will be mostly pipelined with the IO and basically come for free.

The ENOMEM path has not been tested.  It should be.

Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

include/linux/swap.h

@@ -12,6 +12,8 @@
 
 struct notifier_block;
 
+struct bio;
+
 #define SWAP_FLAG_PREFER	0x8000	/* set if swap priority specified */
 #define SWAP_FLAG_PRIO_MASK	0x7fff
 #define SWAP_FLAG_PRIO_SHIFT	0
@@ -216,7 +218,8 @@ extern void swap_unplug_io_fn(struct backing_dev_info *, struct page *);
 /* linux/mm/page_io.c */
 extern int swap_readpage(struct file *, struct page *);
 extern int swap_writepage(struct page *page, struct writeback_control *wbc);
-extern int rw_swap_page_sync(int, swp_entry_t, struct page *);
+extern int rw_swap_page_sync(int rw, swp_entry_t entry, struct page *page,
+				struct bio **bio_chain);
 
 /* linux/mm/swap_state.c */
 extern struct address_space swapper_space;

kernel/power/swap.c

@@ -49,18 +49,16 @@ static int mark_swapfiles(swp_entry_t start)
 {
 	int error;
 
-	rw_swap_page_sync(READ,
-			  swp_entry(root_swap, 0),
-			  virt_to_page((unsigned long)&swsusp_header));
+	rw_swap_page_sync(READ, swp_entry(root_swap, 0),
+			  virt_to_page((unsigned long)&swsusp_header), NULL);
 	if (!memcmp("SWAP-SPACE",swsusp_header.sig, 10) ||
 	    !memcmp("SWAPSPACE2",swsusp_header.sig, 10)) {
 		memcpy(swsusp_header.orig_sig,swsusp_header.sig, 10);
 		memcpy(swsusp_header.sig,SWSUSP_SIG, 10);
 		swsusp_header.image = start;
-		error = rw_swap_page_sync(WRITE,
-					  swp_entry(root_swap, 0),
-					  virt_to_page((unsigned long)
-						       &swsusp_header));
+		error = rw_swap_page_sync(WRITE, swp_entry(root_swap, 0),
+				virt_to_page((unsigned long)&swsusp_header),
+				NULL);
 	} else {
 		pr_debug("swsusp: Partition is not swap space.\n");
 		error = -ENODEV;
@@ -88,16 +86,37 @@ static int swsusp_swap_check(void) /* This is called before saving image */
  *	write_page - Write one page to given swap location.
  *	@buf:		Address we're writing.
  *	@offset:	Offset of the swap page we're writing to.
+ *	@bio_chain:	Link the next write BIO here
  */
 
-static int write_page(void *buf, unsigned long offset)
+static int write_page(void *buf, unsigned long offset, struct bio **bio_chain)
 {
 	swp_entry_t entry;
 	int error = -ENOSPC;
 
 	if (offset) {
+		struct page *page = virt_to_page(buf);
+
+		if (bio_chain) {
+			/*
+			 * Whether or not we successfully allocated a copy page,
+			 * we take a ref on the page here.  It gets undone in
+			 * wait_on_bio_chain().
+			 */
+			struct page *page_copy;
+			page_copy = alloc_page(GFP_ATOMIC);
+			if (page_copy == NULL) {
+				WARN_ON_ONCE(1);
+				bio_chain = NULL;	/* Go synchronous */
+				get_page(page);
+			} else {
+				memcpy(page_address(page_copy),
+					page_address(page), PAGE_SIZE);
+				page = page_copy;
+			}
+		}
 		entry = swp_entry(root_swap, offset);
-		error = rw_swap_page_sync(WRITE, entry, virt_to_page(buf));
+		error = rw_swap_page_sync(WRITE, entry, page, bio_chain);
 	}
 	return error;
 }
@@ -185,37 +204,68 @@ static int get_swap_writer(struct swap_map_handle *handle)
 	return 0;
 }
 
-static int swap_write_page(struct swap_map_handle *handle, void *buf)
+static int wait_on_bio_chain(struct bio **bio_chain)
 {
-	int error;
+	struct bio *bio;
+	struct bio *next_bio;
+	int ret = 0;
+
+	if (bio_chain == NULL)
+		return 0;
+
+	bio = *bio_chain;
+	while (bio) {
+		struct page *page;
+
+		next_bio = bio->bi_private;
+		page = bio->bi_io_vec[0].bv_page;
+		wait_on_page_locked(page);
+		if (!PageUptodate(page) || PageError(page))
+			ret = -EIO;
+		put_page(page);
+		bio_put(bio);
+		bio = next_bio;
+	}
+	*bio_chain = NULL;
+	return ret;
+}
+
+static int swap_write_page(struct swap_map_handle *handle, void *buf,
+				struct bio **bio_chain)
+{
+	int error = 0;
 	unsigned long offset;
 
 	if (!handle->cur)
 		return -EINVAL;
 	offset = alloc_swap_page(root_swap, handle->bitmap);
-	error = write_page(buf, offset);
+	error = write_page(buf, offset, bio_chain);
 	if (error)
 		return error;
 	handle->cur->entries[handle->k++] = offset;
 	if (handle->k >= MAP_PAGE_ENTRIES) {
+		error = wait_on_bio_chain(bio_chain);
+		if (error)
+			goto out;
 		offset = alloc_swap_page(root_swap, handle->bitmap);
 		if (!offset)
 			return -ENOSPC;
 		handle->cur->next_swap = offset;
-		error = write_page(handle->cur, handle->cur_swap);
+		error = write_page(handle->cur, handle->cur_swap, NULL);
 		if (error)
-			return error;
+			goto out;
 		memset(handle->cur, 0, PAGE_SIZE);
 		handle->cur_swap = offset;
 		handle->k = 0;
 	}
-	return 0;
+out:
+	return error;
 }
 
 static int flush_swap_writer(struct swap_map_handle *handle)
 {
 	if (handle->cur && handle->cur_swap)
-		return write_page(handle->cur, handle->cur_swap);
+		return write_page(handle->cur, handle->cur_swap, NULL);
 	else
 		return -EINVAL;
 }
@@ -232,6 +282,8 @@ static int save_image(struct swap_map_handle *handle,
 	int ret;
 	int error = 0;
 	int nr_pages;
+	int err2;
+	struct bio *bio;
 	struct timeval start;
 	struct timeval stop;
 
@@ -240,11 +292,13 @@ static int save_image(struct swap_map_handle *handle,
 	if (!m)
 		m = 1;
 	nr_pages = 0;
+	bio = NULL;
 	do_gettimeofday(&start);
 	do {
 		ret = snapshot_read_next(snapshot, PAGE_SIZE);
 		if (ret > 0) {
-			error = swap_write_page(handle, data_of(*snapshot));
+			error = swap_write_page(handle, data_of(*snapshot),
+						&bio);
 			if (error)
 				break;
 			if (!(nr_pages % m))
@@ -252,7 +306,10 @@ static int save_image(struct swap_map_handle *handle,
 			nr_pages++;
 		}
 	} while (ret > 0);
+	err2 = wait_on_bio_chain(&bio);
 	do_gettimeofday(&stop);
+	if (!error)
+		error = err2;
 	if (!error)
 		printk("\b\b\b\bdone\n");
 	show_speed(&start, &stop, nr_to_write, "Wrote");
@@ -307,7 +364,7 @@ int swsusp_write(void)
 	error = get_swap_writer(&handle);
 	if (!error) {
 		unsigned long start = handle.cur_swap;
-		error = swap_write_page(&handle, header);
+		error = swap_write_page(&handle, header, NULL);
 		if (!error)
 			error = save_image(&handle, &snapshot,
 					header->pages - 1);

mm/page_io.c

@@ -156,10 +156,12 @@ out:
  * We use end_swap_bio_read() even for writes, because it happens to do what
  * we want.
  */
-int rw_swap_page_sync(int rw, swp_entry_t entry, struct page *page)
+int rw_swap_page_sync(int rw, swp_entry_t entry, struct page *page,
+			struct bio **bio_chain)
 {
 	struct bio *bio;
 	int ret = 0;
+	int bio_rw;
 
 	lock_page(page);
 
@@ -170,11 +172,22 @@ int rw_swap_page_sync(int rw, swp_entry_t entry, struct page *page)
 		goto out;
 	}
 
-	submit_bio(rw | (1 << BIO_RW_SYNC), bio);
-	wait_on_page_locked(page);
+	bio_rw = rw;
+	if (!bio_chain)
+		bio_rw |= (1 << BIO_RW_SYNC);
+	if (bio_chain)
+		bio_get(bio);
+	submit_bio(bio_rw, bio);
+	if (bio_chain == NULL) {
+		wait_on_page_locked(page);
 
-	if (!PageUptodate(page) || PageError(page))
-		ret = -EIO;
+		if (!PageUptodate(page) || PageError(page))
+			ret = -EIO;
+	}
+	if (bio_chain) {
+		bio->bi_private = *bio_chain;
+		*bio_chain = bio;
+	}
 out:
 	return ret;
 }